2. Theory
2.3 Maintenance programs
Table 1 – Overview of maintenance programs [31].
Strategy Traits
Corrective Maintenance (CM) • No planning
• Ideal for low priority equipment Preventive Maintenance (PM) • Defined schedules
• Based on statistics
• Ideal for medium priority equipment Predictive Maintenance (PdM) • Based on actual condition of
equipment
• Condition monitoring is required
• Ideal for high priority equipment Proactive Maintenance (Pro-M) • Based on Preventive Maintenance
and Predictive Maintenance
• Focuses on identifying root causes for failure and resolving them
Historically, the maintenance of equipment and systems have evolved alongside the technological development [5]. Today the process of defining the most successful maintenance strategy for each equipment is called Reliablilty-Centered Maintenance (RCM). This strategy is implemented to optimize the maintenance program in regard to the failure modes and criticality analysis. The historical development of maintenance is tied to the technological development, seen from a historical perspective the timeline is illustrated in the figure below.
19 Figure 9 - Historical development of maintenance [5].
2.3.1 Corrective Maintenance (CM)
Corrective maintenance is a technique used when the equipment in question does not have a critical effect on HSE or the further operations of the system. The maintenance is carried out after a failure is occurred and the goal is to bring the asset back to operational function [32].
This philosophy is based upon the belief that the costs accrued by downtime and repairs are lower than the cost of implementing a maintenance program, which leads to CM being only useful for non-critical systems.
20 Table 2 – Overview of Corrective Maintenance
Corrective Maintenance
Pros Cons
• No planning: All downtime is unplanned, and replacements does not need to be available
• Complete use time: The components will ha a maximum lifetime and will not be replaced with useful life remaining.
• Financial loss: Downtime is proportional with loss of production and revenue
• Unstable production: The downtime might cause bad customer relations as they would want their product in due time. The customer is the end point of the value chain and essential to the making of profit.
• Missed learning: No time to implement counter measures to avoid problems in the future.
2.3.2 Preventive Maintenance (PM)
Preventive maintenance is the preferred maintenance program when the cost of maintenance is lower than the cost of a breakdown and the goal of the program is to minimize downtime. This program includes systematic inspection of the asset in order to detect problems and failures in order to correct them. It is based upon historical data and prior knowledge of how and when the components in the asset breaks down. This data combined with a risk and time-based approach is used to schedule inspections and maintenance. The schedule is determined and based on prior breakdown history, component specifications and operator recommendations [33].
21 Table 3 –Overview of Preventive Maintenance
Preventive Maintenance
Pro Cons
• Minimizing of unplanned downtime as parts are changed before failure.
• In general, the life expectancy of critical equipment is increased as the integrity of the asset is maintained.
• Availability of spare parts and
• Cost of maintenance: Scheduling and planning increases the cost related to maintenance. variations in production, and such the inspection and maintenance routines might always not be optimally suited
2.3.3 Predictive Maintenance (PdM)
Predictive Maintenance in regards to the Industry 4.0 paradigm is a program used for preventing failure in assets by monitoring, collecting and analysing data to predict and identify failures before they take place. The main goal of PdM is to maximize uptime and the maintenance strategy is used when the failure of the asset has a critical consequence in regard to operations or HSE.
A key concept in PdM systems is the term Remaining Useful Life (RUL). This term a prediction of the time remaining until a machine failure that requires a repair or replacement.
Maintenance teams are able to use this prediction to evaluate and optimize the maintenance schedule. The inputs needed for establishing this model is mainly dependent on the data available and is mainly indicators of the condition of the machine [28].
22
In general, maintenance is performed when certain indicators regarding the assets health signals that the condition is deteriorating and a certain “threshold” is reached. These indicators are gained through various condition monitoring techniques, ranging from physical inspection to real – time sensor data. Some of the most used condition monitoring techniques are described in chapter 2.4.
One can observe PdM in different levels of maturity and a framework for identifying these where developed in a report by PwC [34]. These four levels can be described as:
Level:1 Visual inspections
Level 2: Instrument inspections and measurements Level 3: Real time condition monitoring
Level 4: Continuous real time condition monitoring where decisions are made with regards to predictive techniques like regression analysis
Where Level 4 is in some cases described as PdM 4.0 or Prescriptive due to its close entanglement with Industry 4.0 [35]. Predictive maintenance is dependent on data, the different levels exist in a Big Data framework and is illustrated in the figure below.
Figure 10 – Framework for Prescriptive maintenance [35].
This latest maturity level revolves around predicting failures and developing the most effective preventive measure by using analytic techniques on big data of elements related to performance of the asset. This kind of predictive maintenance involves exploiting new technological
23
advancements within artificial intelligence, machine learning and pattern recognition as the basis of decision making. For what this thesis is concerned the focus will be on predictive maintenance as a way to utilize a Remaining Useful Lifetime (RUL) model to evaluate the state and remaining lifetime of the machine.
Table 4 – Overview of Predictive Maintenance [35].
Predictive Maintenance service personnel due to scheduling and advance notice of failures.
• Close to complete use time of parts as they are replaced close to the failure time.
• Less expenses as the cost related to unplanned downtime and ineffective maintenance.
• In general, the life expectancy of critical equipment is increased as the integrity of the asset is maintained.
• High costs: Establishing a Condition Monitoring system can involve high capital and operational expenses.
• Less predictable maintenance cost:
The replacement when needed philosophy is less predictable and will give more short-term cost.
• Need for change: The organization will have to make a big transition from a schedule-based program to more flexible real-time solution program.
2.3.4 Proactive Maintenance (Pro-M)
Proactive Maintenance is in many ways a combination of maintenance and continuous improvement. Preventive and predictive maintenance strategies influence proactive maintenance and is in many ways more of an approach than a strategy. The approach is mainly based upon a preventive maintenance system where it rather works toward eliminating the root causes of failure. The proactive approach helps enable organizations to perform maintenance only when it is necessary. It is based upon gathered maintenance data and requires one to address more systematic elements of the maintenance program rather than the machines
24
themselves. The data collection condition monitoring is an important aspect of proactive maintenance programs as they need to track data and trends to predict fault scenarios [36].
The main objectives of the proactive maintenance program are to [36]:
- Identify the root causes for failure
- Resolve potential failure scenarios before they manifest - Extend the Remaining Useful Lifetime (RUL)
Table 5 - Proactive maintenance [35] [36].
Proactive Maintenance service personnel due to scheduling and advance notice of failures.
• Close to complete use time of parts as they are replaced close to the failure time.
• Less expenses as the cost related to unplanned downtime and ineffective maintenance. If optimized, it should provide the lowest maintenance in general due to the improvement of
“weak” links in the organization.
• In general, the life expectancy of critical equipment is increased as the integrity of the asset is maintained.
• The most optimized maintenance program.
• High costs: Establishing a Condition Monitoring system can involve high capital and operational expenses.
• Less predictable maintenance cost:
The replacement when needed philosophy is less predictable and will give more short-term cost.
• Need for change: The organization will have to make a big transition from a schedule-based program to more flexible real-time solution program.
• Difficulty: Out of all the different options proactive maintenance requires the most skill and competency to be effectively implemented.
25